Download Planetary Radii Across Five Orders of Magnitude in Mass and Stellar

The Astrophysical Journal, 659:1661–1672, 2007 April 20
# 2007. The American Astronomical Society. All rights reserved. Printed in U.S.A.
PLANETARY RADII ACROSS FIVE ORDERS OF MAGNITUDE IN MASS
AND STELLAR INSOLATION: APPLICATION TO TRANSITS
J. J. Fortney,1,2,3 M. S. Marley,1 and J. W. Barnes4
Received 2006 November 3; accepted 2006 December 20
ABSTRACT
To aid in the physical interpretation of planetary radii constrained through observations of transiting planets, or eventually direct detections, we compute model radii of pure hydrogen-helium, water, rock, and iron planets, along with various mixtures. Masses ranging from 0.01 Earth masses to 10 Jupiter masses at orbital distances of 0.02–10 AU are
considered. For hydrogen-helium rich planets, our models are the first to couple planetary evolution to stellar irradiation over a wide range of orbital separations (0.02–10 AU ) through a nongray radiative-convective equilibrium
atmosphere model. Stellar irradiation retards the contraction of giant planets, but its effect is not a simple function of
the irradiation level: a planet at 1 AU contracts as slowly as a planet at 0.1 AU. We confirm the assertion of Guillot that
very old giant planets under modest stellar irradiation (like that received by Jupiter and Saturn) develop isothermal
atmospheric radiative zones once the planet’s intrinsic flux drops to a small fraction of the incident flux. For hydrogenhelium planets, we consider cores up to 90% of the total planet mass, comparable to those of Uranus and Neptune. If
‘‘hot Neptunes’’ have maintained their original masses and are not remnants of more massive planets, radii of 0.30–
0.45 RJ are expected. Water planets are 40%–50% larger than rocky planets, independent of mass. Finally, we provide
tables of planetary radii at various ages and compositions, and for ice-rock-iron planets we fit our results to analytic
functions, which will allow for quick composition estimates, given masses and radii, or mass estimates, given only
planetary radii. These results will assist in the interpretation of observations for both the current transiting planet
surveys as well as upcoming space missions, including COROT and Kepler.
Subject headingg
s: binaries: eclipsing — planetary systems
1. INTRODUCTION
and evolution. Tentatively, some work in this general direction is
beginning (Guillot et al. 2006).
The transit method of planet detection is biased toward finding
planets that orbit relatively close to their parent stars. This means
that radial velocity follow-up will be possible for some planets as
the stellar ‘‘wobble’’ signal is larger for shorter period orbits. However, for transiting planets that are low mass, or that orbit very
distant stars, stellar radial velocity measurements may not be possible. For planets at larger orbital distances, radial velocity observations may take years. Therefore, for the foreseeable future a
measurement of planetary radii will be our only window into the
structure of these planets. Estimates of masses will still be important, however, for useful constraints on planet formation theories.
This will have to involve some degree of presumption regarding
the composition of planets. Orbital distances may give some clues
as to a likely composition, but our experience over the past decade
with Pegasi planets (or ‘‘hot Jupiters’’) has shown us the danger of
assuming certain types of planets cannot exist at unexpected orbital distances. Here we compute planetary radii as function of
mass, composition, and stellar irradiation to help in these future
planetary mass and composition estimates.
In x 2 we give a brief overview of the COROT and Kepler
missions. In x 3 we describe the scope of the calculations and our
aims in this study. Later, x 4 describes our methods, while x 5 and
x 6 give our results for ice-rock-iron planets and gaseous planets,
respectively. Finally, x 7 is our conclusions and suggestions for
future work.
We are still in the early days of a revolution in the field of planetary sciences that was triggered by the discovery of planets around
other stars. Exoplanets now number over 200, with masses as small
as 5–7 M (Rivera et al. 2005; Beaulieu et al. 2006). Comparative
planetology, which once included only our solar system’s planets
and moons, now includes sub-Neptune to super-Jupiter-mass planets in other solar systems.
Currently the most important class of exoplanets are those that
transit the disk of their parent stars, allowing for a determination
of planetary radii. The 14 confirmed transiting planets observed
to date are all more massive than Saturn, have orbital periods of
only a few days, and orbit stars bright enough such that radial velocities can be determined, allowing for a calculation of planetary
masses and bulk densities (see Charbonneau et al. 2007a). A planetary mass and radius allows us a window into planetary composition (Guillot 2005). The 14 transiting planets are all gas giants
(see Guillot et al. 1996; Guillot & Showman 2002; Bodenheimer
et al. 2003; Burrows et al. 2003; Baraffe et al. 2005; Laughlin et al.
2005; Arras & Bildsten 2006 for calculations of structure and contraction of many planets) although one planet, HD 149026b,
appears to be 2/3 heavy elements by mass (Sato et al. 2005;
Fortney et al. 2006; Ikoma et al. 2006). Understanding how the
transiting planet mass-radius relations change as a function of orbital distance, stellar mass, stellar metallicity, or UV flux, will provide insight into the fundamentals of planetary formation, migration,
2. UPCOMING SPACE MISSIONS
1
Space Science and Astrobiology Division, NASA Ames Research Center,
Moffett Field, CA 94035; [email protected], [email protected].
2
Spitzer Fellow.
3
Carl Sagan Center, SETI Institute, Mountain View, CA 94043.
4
Lunar and Planetary Laboratory and Department of Planetary Sciences,
University of Arizona, Tucson, AZ 85721; [email protected].
The French /European COROT mission, launched in 2006
December, and the American Kepler mission, set to launch in 2008
November will revolutionize the study of exoplanets. COROT will
monitor 12,000 stars in each of five different fields, each for 150
1661
1662
FORTNEY, MARLEY, & BARNES
continuous days ( Bordé et al. 2003). Planets as small as 2 R
should be detectable around solar-type stars (Moutou et al. 2006).
The mission lifetime is expected to be at least 2.5 yr. The Kepler
mission will continuously monitor one patch of sky, monitoring
over 100,000 main-sequence stars (Basri et al. 2005). The expected
mission lifetime is at least 4 yr. Detection of sub-Earth size planets
is the mission’s goal, with detection of planets with radii as small at
1 Mercury radius possible around M stars. With these missions,
perhaps hundreds of planets will be discovered with masses ranging from sub-Mercury to many times that of Jupiter. Of course,
while planets close to their parent stars will preferentially be found,
due to their shorter orbital periods and greater likelihood to transit, planetary transits will be detected at all orbital separations.
In general, the detection of three successive transits will be
necessary for a confirmed detection, which will limit confirmed
planetary-radius objects to 1.5 AU. It is important to remember
that COROT and Kepler will not determine planetary masses,
only radii, so until follow-up radial velocity or astrometry work
is done, planetary radii will be our only window into the composition of these new planets.
3. FOCUS AND SCOPE OF CALCULATIONS
In the next few years technology will allow the detection of
transiting planets that range from many Jupiter masses (1 MJ ¼
317:89 M ), to perhaps as small as 0.01 M . Interpreting these
observations will require planetary radii to be understood over
more than 5 orders of magnitude in mass. Discoveries to date have
been surprising, including transiting planets that are larger than
standard models can explain (such as HD 209458b and HAT-P-1b),
Saturn-mass planets that have 4 Neptune masses worth of heavy
elements (such as HD 149026b), and among those that do not
transit, Neptune-mass planets that are quite hot and perhaps not
ice-rich (such as HD 68930b), and 5–10 M planets that some
are calling ‘‘super-Earths’’ (such as Gliese 876d ). Therefore, we
think it is useful to take as broad a view as possible. We make few
assumptions regarding composition, and we calculate radii of pure
hydrogen-helium planets as small as 10 M , water and rock dominated planets up to 1000 M (3 MJ ), along with many compositions in between. Since we have the theoretical tools at our
disposal, excluding some compositions a priori is unnecessary at
this point.
We use state-of-the-art equations of state (EOS) for iron (Sesame
2140; Lyon & Johnson 1992), olivine (for generic ‘‘rock’’; ANEOS;
Thompson 1990), water (ANEOS; Thompson 1990), helium
(Saumon et al. 1995), and hydrogen (Saumon et al. 1995). Although detailed models for the thermal evolution of super terrestrial planets have been published (Valencia et al. 2006; Ehrenreich
et al. 2006), issues discussed in such detailed models, such as lithospheric thickness and interior temperature structure have only a
small effect on planetary radii (temperature effects may reach 4%
in radius for the highly irradiated ‘‘Super Mercuries’’ of Valencia
et al. 2006) and are unobservable in the forseeable future. Here
we are interested in the most readily observed quantity: planetary
radii. Even for precisely determined light curves, planetary radii
deteminations are somewhat imprecise owing to uncertainties in
parent star radii, which can often reach 10%.
For hydrogen/ helium planets, irradiation from a parent star is
a significant energy source that strongly affects evolution and contraction, and must be accounted for. We accurately incorporate
stellar heating into our evolution (contraction) calculations through
self-consistent nongray model atmospheres under irradiation from
0.02 to 10 AU. This is the first investigation of the evolution and
contraction of hydrogen/helium planets under irradiation over
such a wide range of orbital separations.
Vol. 659
Other authors have previously computed radii for some subset
of the compositions here. Zapolsky & Salpeter (1969) is perhaps the best known, in which they calculated the radius of zerotemperature spheres of H, He, C, Fe, and Mg for masses from
0.3 M to 10 M. Their Thomas-Fermi-Dirac EOSs are accurate
at pressures where some degree of ionization takes place, but less
so for the lower pressures of our solar system’s terrestrial planets.
Similary, Stevenson (1982) calculated radii from 1 to 1000 M for
cold and warm H, H/He, ice, and rock planets. Valencia et al. (2006)
have recently investigated the radii of Earth-like and Mercury-like
planets up to 10 M . Saumon et al. (1996) have calculated planetary
radii of giant planets with and without cores. Guillot et al. (1996)
also investigated the radii of 51 Peg b-like planets, with compositions ranging from H /He to rock. More recently Bodenheimer
et al. (2003) have calculated radii of H /He planets with and without cores, at various orbital separations, but these models lack
realistic atmospheric boundary conditions and equations of state
for core materials. Our results for water, rock, and iron bodies do
not differ substantially from these other works, but we compute
accurate radii over a wider range of masses, and also include mixed
compositions. However, our calculations for the radii of H/He-rich
planets are an important improvement over other works, as we explicitly and accurately include irradiation from a parent star. In
addition, we include the effects on planetary radii across a wide
range of core masses.
4. METHODS
4.1. Equations of State
First we will look in a bit more detail into our equations of state
before elaborating on the construction of the planet models. The
total pressure, (P), can be expressed as
P ¼ P0 þ Pth ;
ð1Þ
where P0 is the pressure at zero temperature and Pth is the thermal pressure. For the degenerate interiors of massive planets, thermal effects are quite small for high atomic number species. For
instance, zero-temperature equations of state for rock are thought
to be accurate to within 1%–2% for use in Uranus and Neptune
interior models (Hubbard & Macfarlane 1980; Hubbard 1984),
so we make no thermal corrections for rock and iron. For water,
thermal pressure can be important at the 10% level. Hubbard &
Macfarlane (1980) find a relation for the Pth of water that is
suitable for the interior pressure-temperature (P-T ) profiles of
Uranus and Neptune. This relation,
Pth ¼ 3:59 ; 105 T ;
ð2Þ
where Pth is in Mbar, is in g cm3, and T is in K, is relevant for
‘‘typical planetary interior conditions.’’ For 4 g cm3 and
T 5000 K, Pth 0:7 Mbar, or 10% of the total pressure, P. For
planets composed of any fraction of water, we assume that interior temperatures within the ice follow the Uranus/Neptune adiabat
of Guillot (2005), which reaches 550 K at 1 kbar and 4000 K at
1 Mbar, and add the thermal pressure correction of equation (2) at
every P0. This is our EOS of ‘‘warm ice.’’ For the time being we
will ignore the additional complexity of including the EOS of
methane and ammonia, solar system ices that are not as abundant
as water. These molecules condense at colder temperatures, and
it is not clear how abundant these species may be in transiting
planets, which are preferentially found relatively close to their
parent stars. In Figure 1 we show our equations of state for cold
ice, warm ice, rock, and iron. For hydrogen/helium envelopes,
No. 2, 2007
PLANETARY RADII
Fig. 1.—Zero-temperature pressure-density relations for iron ( Fe), rock
( Mg2SiO4), and water ice ( H20). For ice, the dashed curve shows our EOS with
the thermal correction described in x 4.1.
we compute internal adiabats with a helium mass fraction Y ¼ 0:28
and do not include heavy elements. A description of this detailed
H /He EOS can be found in Saumon et al. (1995).
4.2. Planetary Structure and Evolution
The structure of spherically symmetric planets in hydrostatic
equilibrium follow the relations set out below. Equations (3) and
(4) define mass conservation and hydrostatic equilibrium, respectively. Equation (5) defines energy conservation, which is employed
in our evolution calculations of planets with a H /He envelope:
@r
1
¼
;
@m 4r 2 @P
Gm
¼
;
@m
4r 4
@L
@S
¼ T
:
@m
@t
ð3Þ
ð4Þ
ð5Þ
Here r is the radius of a mass shell, m is the mass of a given shell,
is the local mass density, P is the pressure, G is the gravitational
constant, L is the planet’s intrinsic luminosity, T is the temperature,
S is the specific entropy, and t is the time.
For planets composed only of water, rock, or iron, we do not
use equation (5), as we assume a constant radius with age. Given
the small thermal component of the pressure for these materials,
and the expected uncertainty in radius measurements, this assumption is valid. For planets where hydrogen and helium make up an
appreciable mass fraction, following the thermal evolution and
contraction of these planets is essential. We note that in these models we do not include additional interior energy sources such as
tidal dissapation. This may be important for the hot Jupiters. We
also neglect helium phase separation, which will add 1000 km
in radius to cold giant planets at Gyr ages (Fortney & Hubbard
2004). Recall that for the planet with the most precisely determined radius HD 209458b, the 1 radius uncertainty is still 1.9%
(1800 km), due to uncertainties in the stellar parameters (Knutson
et al. 2007).
Our evolution code for the calculation of the cooling and contraction of adiabatic giant planets is well-tested. It has been used
to produce evolutionary models of Jupiter and Saturn (Fortney &
Hubbard 2003), cool extrasolar giant planets (Fortney & Hubbard
1663
2004; Marley et al. 2007a), hot Jupiters (Fortney et al. 2006), and
it is described in detail in Fortney & Hubbard (2003) and Fortney
(2004). For all of these planets, it is the radiative atmosphere that
serves as the bottleneck for cooling above the adiabatic H/He envelope. This is accounted for with our fully nongray, self-consistent
model atmosphere grids. The importance of using detailed atmosphere models for evolutionary calculations of hot Jupiters is
discussed in Baraffe et al. (2003) and Marley et al. (2007a).
Below the H /He envelope we assume that heavy elements are
found within a distinct core. To model this core we use the EOS of
a 50/50 by mass ice/rock mix using the ANEOS zero-temperature
water and olivine EOSs (Thompson 1990). The compositions of
the cores of Jupiter and Saturn are not known, and given that we
model planets that likely formed at a variety of orbital distances, in
which different ratios of ice/rock could be accumulated, this simple choice is a reasonable one. We ignore the heat content of the
core on the thermal evolution of the planets. This is often done for
evolutionary models of Jupiter and Saturn (Hubbard1977; Saumon
et al. 1992; Fortney & Hubbard 2003), as the error involved is small
compared to other unknowns. Please see Fortney et al. (2006) for
additional discussion on this point.
We also neglect the ‘‘transit radius’’ effect: the apparent radius
of a transiting planet is the radius where the slant optical depth
through the planet’s atmosphere reaches unity. The corresponding atmospheric pressure can vary across many orders of magnitude, depending on the wavelength (Hubbard et al. 2001; Fortney
et al. 2003). Burrows et al. (2003) and Baraffe et al. (2003) have
estimated this effect to be 10% and 5%, respectively, for HD
209458b, compared to some reference radius, such as the radiativeconvective boundary or the 1 bar level. The HST light curve of
Brown et al. (2001) and Charbonneau et al. (2002) was obtained in
a narrow wavelength band that overlaps the strong sodium D-line
absorptions at 589 nm. Across a broad visible wavelength band, for
most planets, the transit radius effect would likely be only a few
percent. For the models presented here, the radii correspond to a
pressure of 1 bar. Based on models from Burrows et al. (2007), an
extension of 5–6 atmosphere scale heights from the 1 bar level
is needed to reach the optical transit radius.
4.3. Atmosphere Grids for Hydrogen-Helium Planets
Giant planets have been shown to be fully convective, or nearly
so, beneath their thin radiative atmospheres (for a review, see
Hubbard et al. 2002). The convection is thought to be quite efficient, and hence it is the radiative planetary atmosphere that serves
as the bottleneck for escaping radiation and controls the cooling
and contraction of the interior (Hubbard 1977). As giant planet atmospheres have a number of atomic and molecular absorbers, including water, ammonia, methane, sodium, and potassium, these
atmospheres are far from blackbodies (Burrows et al. 1997; Marley
et al. 1999; Sudarsky et al. 2000). A model atmosphere grid, which
serves as the upper boundary condition in these evolution calculations, relates the specific entropy (S ) of the planet’s internal
adiabat and atmospheric surface gravity (g) to the planet’s effective temperature (TeA ). While S and g are calculated from the planetary structure, TeA can only be accurately determined from a
nongray planetary atmosphere code.
To compute the boundary condition for the evolution, we compute self-consistent radiative-convective equilibrium atmospheric
structure models on a large grid of gravities, intrinsic effective
temperatures, and incident fluxes. Each model computes an atmospheric temperature structure, accounting for deposition and
reradiation of incident starlight and convective transport and emission of internal thermal energy. The model, based on one developed
1664
FORTNEY, MARLEY, & BARNES
Vol. 659
Fig. 2.—Subset of pressure-temperature profiles taken from two of our grids.
On the left are profiles for a planet at 9.5 AU with g ¼ 13 m s2, decreasing in Tint
with values of 60, 50, 40, 30, 20, 10, and 3 K. On the right are profiles for a planet at
0.1 AU with g ¼ 40 m s2, decreasing in TeA with values of 1000, 630, 400, 250,
160, and 100 K. The solid portions of the profiles are radiative regions and the
dashed portions are convective regions.
Fig. 3.—Pressure-temperature profiles for 4.5 Gyr Jupiter-like planets ( g ¼
25 m s2, Tint ¼ 100 K) from 0.02 to 10 AU from the Sun. Distance from the Sun
in AU is color coded along the right side of the plot. Thick lines are convective
regions, while thin lines are radiative regions. The profiles at 5 and 10 AU show
deviations that arise from numerical noise in the chemical equilibrium table near
condensation points, but this has a negligible effect on planetary evolution.
for Titan (McKay et al. 1989), was originally applied to the study of
giant planet thermal structure by Marley & McKay (1999) and has
also been employed to study the atmospheres of the hot Jupiters
(Marley 1998; Fortney et al. 2005, 2006) as well as brown dwarfs
( Marley et al. 1996, 2002; Burrows et al. 1997), albeit without
incident radiation. The radiative transfer methods, chemical equilibrium calculations, and molecular and atomic opacities are summarized in the above publications as well as R. S. Freedman & K.
Lodders (2007, in preparation). We assume solar metallicity atmospheres (Lodders 2003), and while the effect of condensation on
atmospheric composition is included in the chemical equilibrium
calculation (Lodders & Fegley 2002, 2006), we neglect the opacity
of clouds.
Guillot & Showman (2002), Baraffe et al. (2003), Burrows et al.
(2004), and Marley et al. (2007b) have previously discussed the
importance of properly incorporating incident stellar flux into the
hot Jupiter atmosphere grids that serve as the upper boundary
condition for the evolution of these planets, but this is also necessary at greater orbital separations for old and relatively lowmass planets. These planets have comparatively little internal
energy, and their intrinsic effective temperature (Tint ) quickly
falls below its equilibrium temperature (Teq ), which is set entirely
by absorption of stellar flux. Here, by definition
whole. In practice this means that the incident stellar flux is diluted by a factor of 1/4 (see Marley et al. 2007a).
The common approximation described in Hubbard (1977) for
including stellar flux into an atmosphere grid computed for
isolated model atmospheres is only valid when Tint k Teq . Following the Hubbard (1977) prescription to small Tint leads to
planetary radii that reach an asymptotic value governed by their
Bond albedo (Chabrier & Baraffe 2000), and hence overestimates the radii of old, cold planets. We have computed model atmosphere grids across a range of surface gravities, for Tint from
50 to 2000 K (and as low at 10 K at low gravity), including the
proper solar insolation5 for orbital distances of 0.02, 0.045, 0.1,
1.0, and 9.5 AU (which is Saturn’s orbital distance).6 At Gyr
ages, we find Bond albedos of 0.05–0.1 at distances less than
0.1 AU, and higher values of 0.3–0.4 from 1 to 10 AU, where
cooler temperatures prevail and sodium and potassium, which absorb strongly in the optical, have condensed into clouds below the
visible atmosphere (Sudarsky et al. 2003). We have elected to
ignore cloud opacity here, since considerable uncertainties remain
concerning their effect on the atmospheric structure and albedos of
extrasolar giant planets (EGPs). We will pursue this area in more
detail in a later paper that focuses on the evolution of Jupiter and
Saturn. A separate important issue is the opacity in the deep atmosphere at pressures near 1 kbar. While the temperature structure of highly irradiated atmospheres near P 1 kbar is of great
importance for understanding giant planet thermal evolution
(Guillot & Showman 2002; Arras & Bildsten 2006), the opacities at these pressures remain highly uncertain (de Pater et al.
2005).
Figure 3 shows P-T profiles computed for a Jupiter-like planet
from 0.02 to 10 AU from the Sun. The surface gravity g is 25 m s2
and Tint ¼ 100 K in all models. These profiles are meant to roughly
illustrate the atmospheres of Jupiter-like planets at 4.5 Gyr. The
value of Tint ¼ 100 K is very close to Jupiter’s current value
4
4
TeA
¼ Tint
þ Teq4 :
ð6Þ
When Tint is small, incident stellar flux dominates over intrinsic
flux and a deep atmospheric radiative zone grows, similar to a
highly irradiated hot Jupiter. This was previously discussed by
J. B. Pollack in the early 1990s and mentioned in Guillot (1999),
but we believe this is the first time this effect has been explicitly
shown with detailed model atmospheres. Figure 2 shows model
atmospheres computed at 0.1 AU ( highly irradiated ) and 9.5
(modestly irradiated ). In both cases once Tint TTeq , an isothermal region connects the deep interior adiabat to the upper atmosphere, whose structure is governed only by absorption of stellar
flux (see Hubeny et al. 2003). Note that the very low Tint values
in Figure 2 for this Saturn-like planet would only occur after several Hubble times of evolution. For all models we calculate a planetwide average P-T profile that is representative of the planet as a
5
The solar spectrum can be dowloaded at http://rredc.nrel.gov/solar/spectra /
am0/ASTM2000.html.
6
Occasionally at higher gravities and low Tint extrapolation off the grid was
performed as well. This generally only affected highly irradiated core-free lowmass planets.
No. 2, 2007
PLANETARY RADII
1665
(Guillot 2005), and we find that in our cooling calculations that
the model planets reach a Tint of 102–110 K at 4.5 Gyr, which
is only weakly dependent on stellar irradiation.
A deep external radiative zone is found in the most highly irradiated models. For the planets at P0.05 AU convection does
not begin until P > 1 kbar. From 0.1 to 2 AU the deep internal
adiabat for all models begins at 300+ bar, but there is a second,
detached convective zone at pressures close to 1 bar. This detached
convective zone grows at stellar distance increases, and by 3 AU
the convective zones have merged. Only when these convective
zones merge is the interior adiabat cooler as a function of orbital
distance. The models from 0.1 to 2 AU have essentially the same
internal adiabat, meaning the planets would have the same radius
at a given mass. As we will see, a striking consequence of this effect is that stellar irradiation at 2 AU has approximately the same
effect on retarding cooling and contraction as at 0.1 AU, even
though the incident fluxes vary by a factor of 400!
5. RESULTS: ICE-ROCK-IRON PLANETS
5.1. Planetary Radii
It seems likely that planets with masses within an order of
magnitude of the Earth’s mass will be composed primarily of more
refractory species, like the planetary ices, rocks, and iron. Within
our solar system, objects of similar radius can differ by over a factor of 3 in mass, due to compositional differences. A planet with
the radius of Mercury, which is potentially detectable with Kepler,
could indicate a mass of 0.055 M , like Mercury itself, or a mass
of 1/3 this value, like Callisto, which has a radius that differs by
only 30 km. With our equations of state, we are able to explore the
radii of objects with any possible combination of ice, rock, and iron.
In order to keep this task manageable, we have limited our calculations to several illustrative compositions. These include pure ice
and ice/rock mixtures, which could be described as ‘‘water worlds’’
or ‘‘Ocean planets.’’ Such objects in our solar system, like the icy
satellites of the outer planets, generally have small masses. However, Kuchner (2003) and Léger et al. (2004) have pointed out
water-rich objects could reach many Earth masses (perhaps as
failed giant planet cores) and migrate inward to smaller orbital distances. We also consider planets composed of pure rock, rock and
iron mixtures, and pure iron, more similar to our own terrestrial
planets. The ice/rock and rock /iron mixtures are computed for
75/25, 50/50, and 25/75 percentages by mass, with ice always overlaying rock, and rock always overlaying iron.
Our results are shown in Figure 4. Since we make few assumptions regarding what is a reasonable planet, we have computed
radii from masses of 0.01 to 1000 M . For all compositions, the
radii initially grow as M 1/3 , but at larger masses, compression
effects become important. As a greater fraction of the electrons
become pressure ionized, the materials begin to behave more
like a Fermi gas, and there is a flattening of the mass-radius curves
near 1000 M . Eventually the radii shrink as mass increases,
with radii falling with M 1/3 (see Zapolsky & Salpeter 1969).
At the top left of Figure 4 we also show the size of various levels
of uncertainty in planetary mass, as a percentages of a given mass,
from 10% to 200%. For instance, if one could determine the mass
of a 1 M planet to within 50%, even a radius determination accurate to within 0.25 R would lead to considerable ambiguity
concerning composition, ranging from 50/50 ice/rock to pure iron.
The shallow slope of the mass-radius curves below a few M makes
accurate mass determinations especially important for understanding composition. In Table 1 we give the mass and radius for a
subset of these planets. We note that from 1 to 10 M we find
Fig. 4.—Mass (in M ) vs. radius (in km and R ) for planets composed for ice,
rock, and iron. The topmost thick black curve is for pure ‘‘warm’’ water ice. (See
text.) The middle thick curve is for pure rock ( Mg2SiO4 ). The bottommost thick
curve is for pure iron ( Fe). The three black thin curves between pure ice and pure
rock, are from top to bottom, 75% ice/25% rock, 50/50, and 25/75. The inner layer
is rock and the outer layer is ice. The gray dotted lines between rock and pure warm
ice are the same pure ice and ice/rock curves, but for zero-temperature ice. The three
black thin curves between pure rock and iron, are from top to bottom, 75% rock /
25% iron, 50/50, and 25/75. The inner layer is iron and the outer layer is rock. Solar
system objects are open circles. At the upper left we show the horizontal extent of
mass error bars, for any given mass, from 10% to 200%.
excellent agreement between our models and the more detailed
‘‘Super-Earth’’ models of Valencia et al. (2006).
5.2. Validation in the Solar System
On Figure 4 we have also plotted, in open circles, the masses
and radii of solar system planets and moons. These planets can
be used to validate our methods. For instance, detailed models of
the Earth’s interior indicate that the Earth is approximately 33%
iron by mass with a core-mantle boundary at 3480 km (Dziewonski
& Anderson 1981). This composition is readily recovered from
Figure 4, where Earth plots between the 25% and 50% iron curves,
but closer to 25%. Our simple Earth model, with a iron/rock boundary at 3480 km yields a planetary radius within 100 km (1.5%
smaller) of the actual Earth. Given that our model lacks thermal
corrections to EOSs that are found in detailed Earth models, and
that we ignore lower density species such as sulfur that are likely
mixed with iron into the Earth’s core, we regard this agreement
as excellent, and entirely sufficient with regard to the expected
radii uncertainties as measured by transit surveys.
Elsewhere in the solar system, one can see that we recover ice/
rock or rock/iron ratios of other bodies, which are derived by more
complex models. A brief overview of the structure of the terrestrial planets and icy moons is given in de Pater & Lissauer (2001).
Earth’s Moon is composed almost entirely of rock, with a very
small iron core of radius P400 km. Here, the Moon (the leftmost
circle) plots on top of the line for pure rock. Mercury is calculated
to be 60% iron by mass, and with our models Mercury falls between the 50/50 (rock/iron) and 25/75 curves, but again, closer to
50/50, which shows excellent agreement. Titan is calculated to be
composed of 35% ices, and again we find excellent agreement,
as Titan falls between the 50/50 (ice/rock) and 25/75 curves, slightly
1666
FORTNEY, MARLEY, & BARNES
Vol. 659
TABLE 1
Ice-Rock-Iron Planetary Radii
Composition
0.010
0.032
0.1
0.32
1.0
3.16
10.0
31.6
100
316
Ice............................................
50/50 .......................................
Rock ........................................
67/33 ( Earth-like) ...................
50/50 .......................................
Iron..........................................
0.38
0.33
0.25
0.24
0.23
0.19
0.55
0.48
0.37
0.34
0.33
0.27
0.79
0.69
0.54
0.50
0.48
0.39
1.12
0.97
0.77
0.71
0.68
0.55
1.55
1.36
1.08
1.00
0.95
0.77
2.12
1.85
1.48
1.36
1.30
1.04
2.87
2.48
1.97
1.80
1.71
1.36
3.74
3.23
2.54
2.31
2.19
1.72
4.68
4.03
3.14
2.84
2.69
2.09
5.43
4.67
3.64
3.29
3.12
2.42
Notes.—Radii of planets are in R . Column headers are planet masses in M .
closer to 25/75. Uranus and Neptune cannot be composed purely of
ice, although ice likely makes up the bulk of their masses.
5.3. Applications
In addition to Table 1, we have also fit the radii of these planets
to analytic functions that are quadratic in log M and linear in
composition. These are equations (7) for ice/rock planets and (8)
for rock /iron, shown below:
R ¼ ð0:0592imf þ 0:0975Þð log M Þ2
þ ð0:2337imf þ 0:4938Þ log M þ ð0:3102imf þ 0:7932Þ;
ð7Þ
R ¼ ð0:0912rmf þ 0:1603Þ(log M )2
þ (0:3330rmf þ 0:7387) log M þ ð0:4639rmf þ 1:1193Þ:
ð8Þ
Here R is in R and M is in M , while imf is the ice mass fraction
(1.0 for pure ice and 0.0 for pure rock) and rmf is the rock mass
fraction (1.0 for pure rock and 0.0 for pure iron). The fits were
performed for planetary masses from 0.01 to 100 M . For ice/
rock, equation (7) is on average accurate to within 2.5%. For
rock /iron, equation (8) is accurate to 1.5%. Accuracy near 1 M
for both equations is better than 0.5%, although deviations can
reach 10% for ice/rock at 0.01 M . If a mass and radius can be
determined for a given planet these equations will allow for a
quick and reliable composition estimate. However, since radial
velocity or astrometric follow-up for these small planets will be
extremely difficult and time consuming, radii may have to suffice
as a proxy for mass for some time. Given an assumed composition,
such as ‘‘Earth-like,’’ one could assign masses to terrestrial-sized
transiting planets. The distribution of planetary masses versus orbital distance and stellar type could then be compared with planet
formation models.
6. RESULTS: HYDROGEN-HELIUM
DOMINATED PLANETS
6.1. Radii of Gas Giants
Planets around the mass of Uranus and Neptune (15 M ) to
objects as large at 75 MJ can be described by the same cooling
Fig. 5.—Planetary radii as a function of time for masses of 0.1 MJ (32 M , A), 0.3 MJ (B), 1.0 MJ (C), and 3.0 MJ (D). The five curve colors code for the five different orbital
separations from the Sun, shown in (C ). Solid lines indicate models without cores and dash dotted lines indicate models with a core of 25 M .
No. 2, 2007
PLANETARY RADII
1667
Fig. 6.—Planetary radii at 4.5 Gyr as a function of orbital distance from the Sun.
Models are calculated at 0.02, 0.045, 0.1, 1.0, and 9.5 AU. Masses are 0.1, 0.3, 1.0,
and 3.0 MJ . Coreless planets (thin lines) and planets with a core of 25 M of heavy
elements (thick lines) are shown. Note the shape of these radius curves and the
flattening between 0.1 and 1.0 AU. The 0.1 MJ planet with a 25 M core is off the
plot at 0.5 RJ .
theory. In general, planets with larger cores will have smaller
radii, and planets closer to their parent stars will have larger radii
at a given age then planets at larger orbital distances. In Figure 5
we plot the contraction of planets from 0.1 to 3 MJ , as a function
of age at various orbital distances. We also show the effect of a
core of 25 M , the approximate mass of heavy elements within
Jupiter and Saturn (Saumon & Guillot 2004). The first feature to
notice is that, independent of mass, the spread in radii between
0.045 and 1 AU is quite small compared to the large radii at
0.02 AU and small radii at 9.5 AU. As expected, the effect of the
25 M core diminishes with increased planet mass, as the core becomes a relatively smaller fraction of the planet’s mass. The radii
at early ages it quite large, especially for the low-mass planets
under intense stellar irradiation. These hot low-gravity planets
are potentially susceptible to evaporation ( Baraffe et al. 2005;
Hubbard et al. 2007).
As discussed in detail by Marley et al. (2007a) the physical
properties of giant planets at young ages are quite uncertain. The
models presented here do not include a formation mechanism and
are arbitrarily large and hot at very young ages. Marley et al. (2007a)
have found that in their implementation of the core-accretion mechanism of giant planet formation ( Pollack et al. 1996; Hubickyj
et al. 2005) giant planets form with radii than can be several tenths
of a Jupiter radius smaller than one computes with an arbitrarily
large and hot start. These differences may last 10 Myr for a
1 MJ object to hundreds of millions of years for planets of several
Jupiter masses. Therefore, determining the radius of young giant
planets has the potential to elucidate their postformation structure, and give us clues to their formation mechanism. In addition,
a young hot Jupiter would place constraints on planetary migration times ( Burrows et al. 2000).
Radii as a function of orbital distance at 4.5 Gyr are shown in
Figure 6. For the planets from 0.3 (Saturn’s mass) to 3 MJ , the radius curves are nearly flat between 0.1 and 1 AU. These models
predict that irradiation effects will remain important for planets
>0.1 AU from their parent stars. Planets with the same composition, from 0.1 to 2 AU, should have nearly the same radii. This is
a consequence of the atmospheric temperature structures shown in
Figure 3. Planets with cores are smaller, but there radii curves also
flatten from 0.1 to 1 AU.
Fig. 7.—Planetary radii at 4.5 Gyr as a function of mass. Models are calculated
at 0.02, 0.045, 0.1, 1.0, and 9.5 AU and are color coded at the bottom of the plot.
The black curve is for a heavy element planet of half ice and half rock. The group of
five colored curves above the black curve is for planets that are 90% heavy elements.
The next higher set of five colored curves are for planets that are 50% heavy elements.
The next higher set, shown in dotted lines, are 10% heavy elements. The highest set
are for core-free planets of pure H /He. The open circles are solar system planets and
the diamonds are extrasolar planets.
We can also consider the effects of extremely large and small
core masses. One view of our full range of 4.5 Gyr planets is shown
in Figure 7, where radii versus mass is plotted at five orbital separations from 0.02 to 9.5 AU. Coreless planets and planets with
cores that are 10%, 50%, and 90% of their total planet mass are
shown. Note that these models have a core mass mass that is a
constant percentage of planetary mass, meaning that a 50% core
mass for a 10 M planet is 5 M , and for a 5 MJ planet it is
800 M . This plot is meant to show extremes, but is still quite
illustrative. Several real planets are shown as well.
In Figure 7 Uranus and Neptune lie very close to the 90%
heavy elements curves, as expected from more detailed models
( Hubbard et al. 1991; Podolak et al. 1995), although these planets likely contain more ice than rock. The radii of ‘‘hot Neptunes,’’
should be similar to that of Uranus and Neptune, if hot Neptunes
do not suffer the effects of significant mass loss and are 10%
H/He, like Uranus and Neptune. Baraffe et al. (2005, 2006) have
shown that if these planets are remnants of much larger original
planets, their radii may well exceed 1 MJ .
Looking at the solar system’s gas giants, it is clear that for
Jupiter7 at 5.2 AU, 10% heavy elements by mass is a reasonable
estimate as it falls between the 1 AU and 9.5 AU 10% heavy element curves. Saturn, at 9.5 AU, is clearly more enhanced in heavy
elements that Jupiter, and from this plot perhaps 20% heavy
7
Jupiter itself plots below 1 RJ because its mean radius is 2.2% smaller than
its equatorial radius at 1 bar, 71,492 km, which has become the standard ‘‘Jupiter
radius.’’ We note that planetary oblateness will be extremely difficult to determine from transit light curves ( Barnes & Fortney 2003; Seager & Hui 2002).
1668
FORTNEY, MARLEY, & BARNES
Fig. 8.—Planetary radii with various core masses at 500 Myr (A) and 4.5 Gyr (B). Models are calculated at 0.02, 0.045, 0.1, 1.0, and 9.5 AU and are color coded at the
bottom of the plot. The black curve is for a heavy element planet of 50% ice and 50% rock. Models with no core, and core masses of 10, 25, 50, and 100 M are computed.
Labels in thick black text show where curves with a given constant core mass fall on the mass/radius curve for heavy elements. Planets at 0.045, 0.1, and 1.0 AU are similar in
radius at every mass, but planets at 0.02 AU (magenta) are significantly larger, while planets at 9.5 AU ( green) are significantly smaller. Core-free models are the curves that
terminate at low mass at the upper left. The open circles are solar system planets and the diamonds are extrasolar planets.
elements would be estimated. Given that one of the first 10 transiting planets found has a core mass of 60–80 M , and it orbits
a star with a metallicity 2.3 times that of the Sun, it is probably
realistic to expect the occasional planet around metal-rich stars with
core masses of 100 M or more.
We now turn to models computed with a constant core mass
(10, 25, 50, and 100 M ) as function of age, mass, and stellar irradiation. Figures 8a and 8b are dense plots that show the radii of
planets at two ages, 500 Myr and 4.5 Gyr. The contraction of planets by 0.1–0.2 RJ across this factor of 10 in age is clear. Planets
at 0.045 AU (blue), 0.1 AU (orange), and 1.0 AU (red) are very
similar in radius at every mass, so their mass-radius curves trace
similar paths, while planets at 0.02 AU (magenta) are always substantially larger, and planets at 9.5 AU ( green) are always substantially smaller. Tables 2, 3, and 4 contain our calculations of
the radii of these planets at ages of 300 Myr, 1 Gyr, and 4.5 Gyr.
Looking at radii at 4.5 Gyr, we find generally good agreement with
the models of Bodenheimer et al. (2003) for planets k0.69 MJ
(within 0.03 RJ ), but at lower masses we predict significantly
smaller planets with differences than can approach 0.4 RJ for coreless models at 0.02 AU. It is likely that in the very low-mass corefree, highly irradiated corner of phase space that evolution models
are perhaps most uncertain. Evaporation effects may be important
here as well.
6.2. Applications
Extending our calculations to parent stars other than the Sun
should be done with some care. Marley et al. (1999) have shown
that planetary Bond albedos are a function of the spectral type of
the primary star. A planet around a later-type star, with a spectrum
peaking closer to the infrared, will have a greater fraction of its
flux absorbed by a planetary atmosphere, meaning a lower planetary Bond albedo. Although this must surely be accounted for
eventually, given current uncertainties in chemistry and cloud formation in these planets, which leads directly to uncertainties in their
atmospheric absorption and scattering properties, for now we will
ignore this effect. However, the incident stellar fluxes of the currently known transiting planets already differ by a factor of 18,8
so the differing magnitudes of incident radiation cannot be ignored.
If a planet’s orbital distance from its parent star is d, then the distance from the Sun that the planet would have to be to receive this
same flux, d , is given by
d ¼ d ðL =LÞ1=2 ;
ð9Þ
or if one uses a mass-radius relation such as,
L=L ¼ ð M =M Þ ;
ð10Þ
where may typically be 3.5 across the H-R diagram, then
d ¼ d ðM =M Þ=2 :
ð11Þ
For instance, HD 209458b, at 0.046 AU from its G 0V parent star
(L 1:6 L ), would receive that same level of irradiation at
0.036 AU from the Sun.
Figure 8 shows the mass and radius of three transiting planets.
The planet with the lowest mass, HD 149026b, receives an incident flux equivalent to that of a planet at 0.026 AU from the
8
See Frederic Pont’s website at http://obswww.unige.ch / pont/ TRANSITS.htm
for an updated tabulation of transiting planet system parameters.
TABLE 2
Giant Planet Radii at 300 Myr
Distance
(AU )
0.052
0.087
0.15
0.24
0.41
0.68
1.0
1.46
2.44
4.1
6.8
11.3
Core
Mass
17
28
46
77
129
215
318
464
774
1292
2154
3594
0.02.......................
0
10
25
50
100
0
10
25
50
100
0
10
25
50
100
0
10
25
50
100
0
10
25
50
100
...
1.102
2.795
0.801
1.595
0.755
1.504
0.727
0.929
0.565
-
2.326
1.388
0.493
1.522
1.012
0.447
1.395
0.956
0.438
1.325
0.921
0.433
0.951
0.733
0.394
-
1.883
1.465
0.945
1.345
1.091
0.793
1.270
1.035
0.767
1.222
1.004
0.754
0.983
0.847
0.664
-
1.656
1.422
1.133
0.801
1.255
1.124
0.968
0.719
1.197
1.084
0.938
0.702
1.169
1.063
0.923
0.693
1.020
0.939
0.826
0.635
-
1.455
1.349
1.220
1.030
0.669
1.240
1.168
1.071
0.921
0.627
1.202
1.134
1.042
0.899
0.618
1.182
1.116
1.027
0.888
0.613
1.070
1.016
0.942
0.823
0.587
1.378
1.325
1.253
1.144
0.939
1.228
1.185
1.124
1.033
0.863
1.198
1.157
1.099
1.011
0.847
1.182
1.141
1.085
0.999
0.839
1.106
1.072
1.024
0.951
0.810
1.342
1.311
1.267
1.193
1.055
1.212
1.185
1.147
1.084
0.968
1.187
1.160
1.123
1.063
0.950
1.173
1.146
1.110
1.051
0.941
1.127
1.104
1.073
1.020
0.920
1.327
1.306
1.275
1.226
1.128
1.206
1.188
1.161
1.119
1.036
1.182
1.164
1.138
1.098
1.018
1.169
1.152
1.127
1.087
1.009
1.146
1.131
1.146
1.072
0.999
1.308
1.295
1.276
1.245
1.187
1.199
1.188
1.173
1.148
1.101
1.178
1.168
1.153
1.129
1.084
1.168
1.158
1.143
1.120
1.075
1.167
1.157
1.142
1.119
1.072
1.311
1.304
1.294
1.276
1.242
1.210
1.204
1.195
1.179
1.148
1.189
1.183
1.174
1.158
1.128
1.179
1.173
1.164
1.149
1.119
1.169
1.163
1.153
1.137
1.107
1.315
1.310
1.304
1.292
1.270
1.203
1.199
1.193
1.183
1.163
1.178
1.174
1.169
1.159
1.140
1.169
1.165
1.159
1.149
1.131
1.156
1.152
1.146
1.137
1.119
1.284
1.281
1.277
1.270
1.256
1.170
1.168
1.164
1.157
1.146
1.144
1.142
1.138
1.132
1.121
1.136
1.134
1.130
1.124
1.113
1.130
1.127
1.124
1.118
1.107
0.045.....................
0.1.........................
1.0.........................
9.5.........................
Notes.—Radii of planets are in RJ . Row 1 column headers are planet masses in MJ , while row 2 is in M . The symbol ‘‘-’’ indicates that the planet mass is smaller
than the given core mass.
TABLE 3
Giant Planet Radii at 1 Gyr
Distance
(AU )
0.052
0.087
0.15
0.24
0.41
0.68
1.0
1.46
2.44
4.1
6.8
11.3
Core
Mass
17
28
46
77
129
215
318
464
774
1292
2154
3594
0.02..........................
0
10
25
50
100
0
10
25
50
100
0
10
25
50
100
0
10
25
50
100
0
10
25
50
100
...
0.909
1.490
0.698
1.298
0.665
1.229
0.646
0.857
0.532
-
1.770
1.150
0.461
1.271
0.888
0.426
1.197
0.847
0.420
1.148
0.823
0.416
0.877
0.683
0.386
-
1.539
1.221
0.838
1.183
0.975
0.739
1.127
0.934
0.719
1.095
0.915
0.709
0.910
0.791
0.631
-
1.387
1.211
1.022
0.746
1.144
1.043
0.908
0.684
1.105
1.012
0.883
0.670
1.086
0.996
0.871
0.663
0.955
0.882
0.780
0.610
-
1.309
1.228
1.121
0.958
0.640
1.163
1.099
1.012
0.877
0.607
1.133
1.072
0.989
0.859
0.600
1.118
1.058
0.977
0.850
0.595
1.003
0.955
0.888
0.784
0.570
1.281
1.234
1.169
1.072
0.888
1.167
1.127
1.072
0.988
0.831
1.143
1.105
1.051
0.970
0.818
1.130
1.092
1.040
0.961
0.811
1.044
1.013
0.970
0.904
0.775
1.258
1.229
1.189
1.122
0.997
1.160
1.134
1.099
1.041
0.932
1.139
1.114
1.080
1.023
0.918
1.128
1.103
1.069
1.014
0.910
1.068
1.047
1.018
0.970
0.878
1.248
1.229
1.200
1.156
1.068
1.157
1.140
1.115
1.077
0.999
1.138
1.122
1.097
1.059
0.984
1.127
1.111
1.087
1.050
0.976
1.089
1.075
1.089
1.021
0.954
1.235
1.224
1.206
1.180
1.130
1.156
1.147
1.132
1.109
1.065
1.139
1.130
1.116
1.094
1.050
1.130
1.121
1.107
1.085
1.042
1.113
1.104
1.090
1.068
1.026
1.244
1.237
1.228
1.211
1.179
1.164
1.158
1.149
1.134
1.105
1.147
1.141
1.132
1.117
1.088
1.137
1.131
1.123
1.108
1.080
1.119
1.113
1.105
1.090
1.063
1.240
1.235
1.229
1.218
1.198
1.149
1.145
1.140
1.130
1.111
1.130
1.126
1.121
1.111
1.093
1.121
1.117
1.112
1.102
1.085
1.109
1.105
1.100
1.091
1.074
1.199
1.197
1.192
1.186
1.173
1.107
1.105
1.101
1.095
1.084
1.087
1.085
1.081
1.076
1.065
1.079
1.077
1.073
1.068
1.057
1.074
1.072
1.069
1.063
1.053
0.045........................
0.1............................
1.0............................
9.5............................
Notes.—Radii of planets are in RJ . Row 1 column headers are planet masses in MJ , while row 2 is in M . The symbol ‘‘-’’ indicates that the planet mass is smaller
than the given core mass.
1670
FORTNEY, MARLEY, & BARNES
Vol. 659
TABLE 4
Giant Planet Radii at 4.5 Gyr
Distance
(AU )
0.052
0.087
0.15
0.24
0.41
0.68
1.0
1.46
2.44
4.1
6.8
11.3
Core
Mass
17
28
46
77
129
215
318
464
774
1292
2154
3594
0.02..........................
0
10
25
50
100
0
10
25
50
100
0
10
25
50
100
0
10
25
50
100
0
10
25
50
100
...
0.726
1.103
0.599
1.068
0.592
1.014
0.576
0.798
0.508
-
1.355
0.934
0.430
1.065
0.775
0.403
1.027
0.755
0.404
0.993
0.738
0.400
0.827
0.653
0.378
-
1.252
1.019
0.756
1.038
0.878
0.686
1.005
0.858
0.675
0.983
0.845
0.666
0.866
0.759
0.611
-
1.183
1.072
0.928
0.695
1.049
0.964
0.846
0.648
1.024
0.942
0.829
0.639
1.011
0.931
0.820
0.633
0.913
0.844
0.750
0.594
-
1.190
1.123
1.032
0.891
0.613
1.086
1.029
0.952
0.831
0.587
1.062
1.008
0.934
0.817
0.582
1.050
0.997
0.924
0.810
0.578
0.957
0.911
0.849
0.754
0.558
1.189
1.148
1.091
1.004
0.841
1.105
1.069
1.019
0.942
0.798
1.085
1.051
1.002
0.928
0.788
1.074
1.041
0.993
0.920
0.782
0.994
0.966
0.926
0.865
0.746
1.179
1.153
1.116
1.056
0.944
1.107
1.083
1.050
0.996
0.896
1.090
1.067
1.034
0.982
0.884
1.081
1.058
1.026
0.974
0.878
1.019
0.999
0.972
0.926
0.842
1.174
1.157
1.131
1.091
1.011
1.108
1.092
1.069
1.033
0.961
1.092
1.077
1.054
1.019
0.949
1.084
1.068
1.046
1.011
0.942
1.037
1.024
1.037
0.973
0.911
1.170
1.160
1.145
1.121
1.076
1.113
1.104
1.090
1.068
1.026
1.099
1.090
1.077
1.055
1.014
1.091
1.082
1.069
1.048
1.007
1.056
1.048
1.035
1.015
0.976
1.178
1.172
1.163
1.148
1.118
1.118
1.112
1.104
1.090
1.062
1.104
1.098
1.090
1.076
1.049
1.096
1.090
1.082
1.068
1.041
1.062
1.057
1.050
1.037
1.012
1.164
1.160
1.155
1.144
1.125
1.099
1.095
1.090
1.081
1.063
1.084
1.080
1.075
1.066
1.049
1.075
1.072
1.067
1.058
1.041
1.055
1.052
1.047
1.039
1.023
1.118
1.116
1.112
1.106
1.095
1.053
1.050
1.047
1.042
1.032
1.038
1.036
1.033
1.027
1.018
1.030
1.028
1.025
1.020
1.010
1.023
1.021
1.018
1.013
1.004
0.045........................
0.1............................
1.0............................
9.5............................
Notes.—Radii of planets are in RJ . Row 1 column headers are planet masses in MJ , while row 2 is in M . The symbol ‘‘-’’ indicates that the planet mass is smaller than
the given core mass. Given the albedos we calculate (see x 4.3) the approximate Teq value at each distance as 1960 K (0.02 AU ), 1300 K (0.045 AU ), 875 K (0.1 AU ),
260 K (1 AU ), and 78 K (9.5 AU ).
Sun. The figure clearly suggests that the planet has a core of perhaps 70 M , which is confirmed by more detailed models (Fortney
et al. 2006). The middle diamond is TrES-1, which receives insolation equal to what is received from the Sun at 0.059 AU. Using these models, one would estimate that the planet has a core of
perhaps 25 M , very similar to the bulk abundance of heavy elements in Jupiter and Saturn (Saumon & Guillot 2004). HD 209458b
is anomalous on this plot and must have an additional interior energy source. However, a 0.1 MJ planet of pure H / He at 0.02 AU
(if such a planet exists) with a similar radius would have the correct radius.
It should also be pointed out that all the hot Jupiters could have
additional energy sources of similar magnitude. As discussed in
Fortney et al. (2006) and Guillot et al. (2006), this would require
planets such as TrES-1 and HD 149026b to have even larger cores.
With a common energy source, differences in the radii of giant
planets would then be attributable to differing masses of heavy
elements within their interiors. This idea is strengthened by the
recent discovery of hot Jupiters with large radii, such as HAT-P-1b
(Bakos et al. 2006) and WASP-1b (Collier Cameron et al. 2007;
Charbonneau et al. 2007b), which shows that inflated radii are
common features of these irradiated planets.
7. DISCUSSION AND CONCLUSIONS
In this paper we have taken a broad look at the radii of planets
at nearly all possible masses and orbital separations. These calculations serve as a baseline for comparisons with transiting planets currently being discovered from the ground and those that will
soon be discovered from space. Since our main effort here is to
calculate radii over a very large phase space, there are surely particular corners of space where these calculations could be improved,
as we outline below. If in the coming years it becomes clear under
what conditions planets may have an additional interior energy
source, and how this changes as a function of orbital separation,
incident stellar flux, and stellar type, this could be included in a new
grid of calculations. If evaporation effects are observed for numerous transiting planets, across a wide range of planetary surface
gravities and orbital separations, then the effects of appreciable
mass loss could be added to these evolution calculations. The work
of Baraffe et al. (2006) and Hubbard et al. (2007) is moving in this
direction.
Ice-rich terrestrial planets close to their parent stars could have
steam atmospheres (Kuchner 2003), which, depending on the atmospheric scale height, could lead to larger radii at a given mass
for these objects. If the cooling of Uranus and Neptune are eventually better understood, then the additional energy due to the cooling of their heavy element interiors could be added to our evolution
models of these lower mass EGPs. Including the EOS of other
forms of rock, or other astrophysical ices, such at CH 4 and NH3,
would be useful when more detailed models are eventually needed.
One could also envision more exotic planets, such as those that
formed in solar systems where C/O > 1, which could lead to carbon dominated ‘‘terrestrial planets’’ ( Kuchner & Seager 2005),
which would likely have radii intermediate between pure rock
and ice (Zapolsky & Salpeter, 1969; M. Kuchner 2006, personal
communication).
As there is in the solar system, there will always be ambiguity
in the bulk composition of exoplanets, especially those that do not
No. 2, 2007
PLANETARY RADII
posses substantial low-density gaseous envelopes. For instance,
since a mixture of H/He gas and rocks has an EOS that is similar
to that of ice, this leads to considerable uncertainty in the composition of Uranus and Neptune. For a terrestrial-mass ‘‘ocean planet,’’
perhaps the best that we can expect will be be an understanding of a
planet’s ice/rock ratio, with a given uncertainty based on the mass
and radius measurements, along with the knowledge that a the
uncertainty could be even larger if the planet has a substantial
iron core. Since iron and silicates condense at similar temperatures
(Lodders 2003), it is possible that these species will condense out
in ratios similar to that found in the Earth, although subsequent
collisions (e.g., as experienced by the Moon and probably Mercury)
could alter this ratio, especially for smaller bodies. Experience will
tell us how to best classify newly discovered terrestrial-type planets,
but at this point we would advocate composition classes based on
simple ratios of ice/rock and rock/iron. Our analytic fits reproduce
the behavior of our models and will allow for mass estimates prior
to radial velocity or astrometric follow-up work.
1671
A clear prediction from our models for giant planets is that irradiation effects on planetary radii are not a simple function of
stellar insolation. At a given planetary mass and composition, we
predict a flattening in planetary radius as a function of orbital distance from 0.1 to 1–2 AU. However, if unknown additional
energy sources are still in play at these orbital distances, then this
effect may be more difficult to see.
Given the recent flurry of transit detections it is likely that we
are now just seeing the tip of the iceberg of transiting planets.
Even before detections from COROT and Kepler are announced
it seems likely that we will have a steady stream of new planets to
challenge our understanding.
We thank Kevin Zahnle for a helpful discussion at the start of
this project. J. J. F. acknowledges the support of a Spitzer Fellowship from NASA and NSF grant AST-0607489 and M. S. M. from
the NASA Origins and Planetary Atmospheres Programs.
REFERENCES
Arras, P., & Bildsten, L. 2006, ApJ, 650, 394
Guillot, T., Santos, N. C., Pont, F., Iro, N., Melo, C., & Ribas, I. 2006, A&A,
Bakos, G. A., et al. 2007, ApJ, 656, 552
453, L21
Baraffe, I., Alibert, Y., Chabrier, G., & Benz, W. 2006, A&A, 450, 1221
Guillot, T., & Showman, A. P. 2002, A&A, 385, 156
Baraffe, I., Chabrier, G., Barman, T. S., Allard, F., & Hauschildt, P. H. 2003,
Hubbard, W. B. 1977, Icarus, 30, 305
A&A, 402, 701
———. 1984, Planetary Interiors ( New York: Van Nostrand Reinhold )
Baraffe, I., Chabrier, G., Barman, T. S., Selsis, F., Allard, F., & Hauschildt, P. H.
Hubbard, W. B., Burrows, A., & Lunine, J. I. 2002, ARA&A, 40, 103
2005, A&A, 436, L47
Hubbard, W. B., Fortney, J. J., Lunine, J. I., Burrows, A., Sudarsky, D., &
Barnes, J. W., & Fortney, J. J. 2003, ApJ, 588, 545
Pinto, P. 2001, ApJ, 560, 413
Basri, G., Borucki, W. J., & Koch, D. 2005, NewA Rev., 49, 478
Hubbard, W. B., Hattori, M. F., Burrows, A., Hubeny, I., & Sudarsky, D. 2007,
Beaulieu, J.-P., et al. 2006, Nature, 439, 437
Icarus, 187, 358
Bodenheimer, P., Laughlin, G., & Lin, D. N. C. 2003, ApJ, 592, 555
Hubbard, W. B., & Macfarlane, J. J. 1980, J. Geophys. Res., 85, 225
Bordé, P., Rouan, D., & Léger, A. 2003, A&A, 405, 1137
Hubbard, W. B., Nellis, W. J., Mitchell, A. C., Holmes, N. C., McCandless, P. C.,
Brown, T. M., Charbonneau, D., Gilliland, R. L., Noyes, R. W., & Burrows, A.
& Limaye, S. S. 1991, Science, 253, 648
2001, ApJ, 552, 699
Hubeny, I., Burrows, A., & Sudarsky, D. 2003, ApJ, 594, 1011
Burrows, A., Guillot, T., Hubbard, W. B., Marley, M. S., Saumon, D., Lunine,
Hubickyj, O., Bodenheimer, P., & Lissauer, J. J. 2005, Icarus, 179, 415
J. I., & Sudarsky, D. 2000, ApJ, 534, L97
Ikoma, M., Guillot, T., Genda, H., Tanigawa, T., & Ida, S. 2006, ApJ, 650, 1150
Burrows, A., Hubney, I., Budaj, J., & Hubbard, W. B. 2007, ApJ, submitted
Knutson, H., Charbonneau, D., Noyes, R. W., Brown, T. M., & Gilliland, R. L.
(astro-ph /0612703)
2007, ApJ, 655, 564
Burrows, A., Hubney, I., Hubbard, W. B., Sudarsky, D., & Fortney, J. J. 2004,
Kuchner, M. J. 2003, ApJ, 596, L105
ApJ, 610, L53
Kuchner, M. J., & Seager, S. 2005, preprint (astro-ph /0504214)
Burrows, A., Marley, M., Hubbard, W. B., Lunine, J. I., Guillot, T., Saumon,
Laughlin, G., Wolf, A., Vanmunster, T., Bodenheimer, P., Fischer, D., Marcy,
D., Freedman, R., Sudarsky, D., & Sharp, C. 1997, ApJ, 491, 856
G., Butler, P., & Vogt, S. 2005, ApJ, 621, 1072
Burrows, A., Sudarsky, D., & Hubbard, W. B. 2003, ApJ, 594, 545
Léger, A., et al. 2004, Icarus, 169, 499
Chabrier, G., & Baraffe, I. 2000, ARA&A, 38, 337
Lodders, K. 2003, ApJ, 591, 1220
Charbonneau, D., Brown, T. M., Burrows, A., & Laughlin, G. 2007a, in ProLodders, K., & Fegley, B. 2002, Icarus, 155, 393
tostars and Planets V, ed. V. B. Reipurth, D. Jewitt, & K. Keil ( Tucson: Univ.
———. 2006, Astrophysics Update 2, ed. J. W. Mason (Chichester: Praxis),
Arizona Press), 701
1–28
Charbonneau, D., Brown, T. M., Noyes, R. W., & Gilliland, R. L. 2002, ApJ,
Lyon, S. P., & Johnson, J. D. 1992, LANL Rep. LA-UR-92-3407 ( Los Alamos:
568, 377
LANL)
Charbonneau, D., Winn, J. N., Everett, M. E., Latham, D. W., Holman, M. J.,
Marley, M. S. 1998, in in ASP Conf. Ser. 134, Brown Dwarfs and Extrasolar
Esquerdo, G. A., & O’Donovan, F. T. 2007b, ApJ, 658, 1322
Planets, ed. R. Rebolo, E. L. Martin, & M. R. Zapatero Osorio (San FranCollier Cameron, A., et al. 2007, MNRAS, 375, 951
cisco: ASP), 383
de Pater, I., Deboer, D., Marley, M., Freedman, R., & Young, R. 2005, Icarus,
Marley, M. S., Fortney, J. J., Hubickyj, O., Bodenheimer, P., & Lissauer, J. J.
173, 425
2007a, ApJ, 655, 541
de Pater, I., & Lissauer, J. J. 2001, Planetary Sciences (Cambridge: Cambridge
Marley, M. S., Fortney, J., Seager, S., & Barman, T. 2007b, in Protostars and
Univ. Press)
Planets V, ed. V. B. Reipurth, D. Jewitt, & K. Keil ( Tucson: Univ. Arizona
Dziewonski, A. M., & Anderson, D. L. 1981, Phys. Earth Planet Inter., 25, 297
Press), 733
Ehrenreich, D., Lecavelier Des Etangs, A., Beaulieu, J.-P., & Grasset, O. 2006,
Marley, M. S., Gelino, C., Stephens, D., Lunine, J. I., & Freedman, R. 1999,
ApJ, 651, 535
ApJ, 513, 879
Fortney, J. J. 2004, Ph.D. thesis, Univ. Arizona
Marley, M. S., & McKay, C. P. 1999, Icarus, 138, 268
Fortney, J. J., & Hubbard, W. B. 2003, Icarus, 164, 228
Marley, M. S., Saumon, D., Guillot, T., Freedman, R. S., Hubbard, W. B.,
———. 2004, ApJ, 608, 1039
Burrows, A., & Lunine, J. I. 1996, Science, 272, 1919
Fortney, J. J., Marley, M. S., Lodders, K., Saumon, D., & Freedman, R. 2005,
Marley, M. S., Seager, S., Saumon, D., Lodders, K., Ackerman, A. S.,
ApJ, 627, L69
Freedman, R. S., & Fan, X. 2002, ApJ, 568, 335
Fortney, J. J., Saumon, D., Marley, M. S., Lodders, K., & Freedman, R. S.
McKay, C. P., Pollack, J. B., & Courtin, R. 1989, Icarus, 80, 23
2006, ApJ, 642, 495
Moutou, C., et al. 2006, in Tenth Anniversary of 51 Peg-b: Status of and
Fortney, J. J., Sudarsky, D., Hubeny, I., Cooper, C. S., Hubbard, W. B., Burrows, A.,
prospects for hot Jupiter studies, ed. L. Arnold, F. Bouchy, & C. Moutou, 342
& Lunine, J. I. 2003, ApJ, 589, 615
Podolak, M., Weizman, A., & Marley, M. 1995, Planet. Space Sci., 43, 1517
Guillot, T. 1999, Science, 286, 72
Pollack, J. B., Hubickyj, O., Bodenheimer, P., Lissauer, J. J., Podolak, M., &
———. 2005, Annu. Rev. of Earth and Planet. Sci., 33, 493
Greenzweig, Y. 1996, Icarus, 124, 62
Guillot, T., Burrows, A., Hubbard, W. B., Lunine, J. I., & Saumon, D. 1996,
Rivera, E. J., et al. 2005, ApJ, 634, 625
ApJ, 459, L35
Sato, B., et al. 2005, ApJ, 633, 465
1672
FORTNEY, MARLEY, & BARNES
Saumon, D., Chabrier, G., & van Horn, H. M. 1995, ApJS, 99, 713
Saumon, D., & Guillot, T. 2004, ApJ, 609, 1170
Saumon, D., Hubbard, W. B., Burrows, A., Guillot, T., Lunine, J. I., &
Chabrier, G. 1996, ApJ, 460, 993
Saumon, D., Hubbard, W. B., Chabrier, G., & van Horn, H. M. 1992, ApJ, 391,
827
Seager, S., & Hui, L. 2002, ApJ, 574, 1004
Stevenson, D. J. 1982, Planet. Space Sci., 30, 755
Sudarsky, D., Burrows, A., & Hubeny, I. 2003, ApJ, 588, 1121
Sudarsky, D., Burrows, A., & Pinto, P. 2000, ApJ, 538, 885
Thompson, S. L. 1990, ANEOS—Analytic Equations of State for Shock
Physics Codes (Sandia Natl. Lab. Doc. SAND89-2951)
Valencia, D., O’Connell, R. J., & Sasselov, D. 2006, Icarus, 181, 545
Zapolsky, H. S., & Salpeter, E. E. 1969, ApJ, 158, 809